attacking the stock market

A lot of smart people have written a lot of software to make money in stocks, which means that, most likely, all the obvious algorithms are farmed out and therefore mostly useless. But I wonder how hard it would be to write a program that could consistently lose money on the stock market?

For instance, if I wrote a program to emulate an overly-exuberant day trader, or to generally just bet that current trends will continue, would it make less than a 50% return? And if so, could I make money by doing the opposite of what it told me?


  1. I imagine you could do pretty well by buying companies going into bankruptcy, unlisted stocks, etc, and not diversifying at all.

    You probably would not be able to make money doing the opposite, since a) you would have to sell short, which might not even be permitted; b) you probably would have trouble finding a buyer at all.

    Another way to lose a ton of money is to conduct transactions at an arbitrarily high rate and fritter money away on brokerage fees, which is obviously non-reversible.

  2. Maybe the trick is to figure out which stocks will go down, and then buy every other stock.

  3. Suppose your program says to buy and sell a few random securities, many times a day (as Sam said, it would fritter money on brokerage fees and be a sure-fired way to lose money).

    What would "doing the opposite of that" be? Buying the whole market and holding it for a long time. Oh wait, that's the strategy advocated by believers of the efficient market hypothesis, right?

  4. So an believer of the efficient market hypothesis and his young daughter are going for a walk when she sees a five dollar bill on the ground.

    "Dad," she says, "can I pick up that money?"

    "No dear," he responds, "if that were a real five dollar bill, someone would have picked it up already."

    Seriously though, yeah, long-term investment in index funds.

  5. The efficient markets hypothesis has sort of always been my personal gremlin, and I even started this post out with a form of it. But I ultimately think that it's just not true, or rather, it's only true as the size of the market approaches infinity and the information distribution approaches uniform. Sure, the average investor can't beat the market, but Warren Buffet can and has. Lots of people make lots of money in stocks, and I'm not prepared to call it all luck. The trick is whether you think you're that guy, and whether you're right. Meh.